IBIS Macromodel Task Group Meeting date: 10 March 2015 Members (asterisk for those attending): Altera: David Banas ANSYS: * Dan Dvorscak * Curtis Clark Avago (LSI) Xingdong Dai Cadence Design Systems: * Ambrish Varma Brad Brim Kumar Keshavan Ken Willis Ericsson: Anders Ekholm IBM Steve Parker Intel: * Michael Mirmak Keysight Technologies: * Fangyi Rao * Radek Biernacki * Nicholas Tzou Maxim Integrated Products: Hassan Rafat Mentor Graphics: John Angulo * Arpad Muranyi Micron Technology: * Randy Wolff Justin Butterfield QLogic Corp. James Zhou Andy Joy eASIC Marc Kowalski SiSoft: Walter Katz Todd Westerhoff * Mike LaBonte Synopsys Rita Horner Teraspeed Consulting Group: Scott McMorrow Teraspeed Labs: * Bob Ross (Note: Agilent has changed to Keysight) The meeting was led by Arpad Muranyi. ------------------------------------------------------------------------ Opens: - Bob: I have a question to ask. - Nicholas Tzou introduced himself. He is working on AMI models for Keysight. -------------------------- Call for patent disclosure: - None ------------- Review of ARs: - Michael M send AMI Direction presentation to Mike L for posting - Done but not posted yet. - Randy send whiteboard circuit diagram to Mike for posting. - Done. - Walter and Randy produce C_comp BIRD. - Randy: Some progress, but we need to resolve issues about parameters. - Arpad to review IBIS specification for min max issues. - In progress. ------------- New Discussion: Question about BIRD 155.2 - Bob: This adds a new Usage "Dep". - Does Resolve_Exists have to be True for this to be legal? - If not then Dep would be interpreted as Out when it is False. - Radek: The editorial committee might handle this. - Bob: Alternatively Dep might be illegal if Resolve_Exists is not True. - Mike L: I would prefer that. - A Dep is suspicious if there is no Resolve_Exists. - Fangyi: Not sure if AMI_Init uses this. - Bob: Should the redriver flow be resolved before this is approved? - Arpad: Walter found a flaw, that should be removed. - Fangyi: It's only a problem when the RX has no AMI_GetWave. - The flow does not mention the new parameter. - Bob: Not sure if the draft BIRD has a Reserved_Parameter. - Radek: It is likely to be incomplete. - Arpad: Walter's point was that it doesn't address the filtering in the first channel passing on to the second channel. - Radek: There is a simpler way to address that. - Arpad: Should that be done for the next IBIS release? - An editorial committee is being formed. - We have 9 approved BIRDs. - Radek: The redriver flow BIRD is posted as 166. - Arpad: It is not yet approved. Should it be in the next IBIS? - Fangyi: It is not that important. - There are better solutions. - Bob: It has no Reserved_Parameters. - Fangyi: It has a parameter, but the flow doesn't use it. - Arpad: If there is a better way that should be worked on. - Should there be a new BIRD if it can't be worked out with Walter? - Fangyi: I sent feedback to Walter but haven't heard back. - I could explain my thinking. - Arpad: The action item has existed for months. - Radek: A competing BIRD might be good. AMI Directionality: - Michael M: We may not have had enough discussion on this. - There has not been much offline feedback. Advance C_comp: - Randy showed a figure drawn after the last meeting. - Randy: The proposal is to add a new node to the C_comp model. where an arrow is shown. - The input buffer is actually somewhat isolated from the pad. - Arpad: Are the buffer triangles independent? - Do they each have C_comp? - Randy: They are circuit elements, not IBIS buffers. - Radek: The whole circuit is present at all times? - Randy: Yes. - Radek: This is dissection a pin into separate in and out portions. - Randy: Currently we can only look at the pad. - Arpad: Would this be in the [Model] keyword? - Randy: It would be part of the new C_comp proposal. - We have Si_Location and Timing_Location, this could be similar. - Bob: C_comp won't solve the problem without having driver and receiver modes. - This is not like location. - Arpad: Where are the clamps? - Randy: They are at the A_signal, the pad. - Arpad: The bottom buffer triangle is just a measurement point. - Randy: Right. - It could be replaced with another symbol. - Bob: Large RC isolates the receiver clamps. - Arpad: Only Vinl Vinh are at the receiver. - Randy: This is a node of interest for signal quality. - Arpad: This could be done easily if there is no second [Model*] keyword. - Michael M showed IBIS-AMI and Direction Indication. - Michael M: This is about models looking forward. - Chapter 10 has something about single ended model support. - slide 5: - Michael M: - The original proposal was to add two Reserved_Parameters. - IBIS Model_type should correspond to AMI_Model_Type. - Model_Direction is the state at any time. - The user controls direction for any given simulation. - Do we want to enable/disable buffers to be I/O? - Or do we want to limit each buffer to one direction. - Arpad: There should be no limitation. - Fangyi: I would prefer separate models. - Walter brought that up. - Michael M: Radek brought it up earlier. - Radek: It's hard to say what is preferable for the model maker. - Ambrish: We can surely have separate .ami files. - Arpad: Could there be multiple root names? - Bob: Only one root name can have Reserved_Parameters. - Michael M: There should be multiple .ami files with associated roots. - We might end up rewriting much of the specification for this. - Arpad: How do you give 2 AMI files for one DLL? - Bob: They can be on separate lines. - One would be for the TX model and one for the RX model. - slide 8 - Michael M: This shows that. - We used to argue against restricting to one AMI file. - Checking would have to be enforced. - Option 2 permits multiple [Algorithmic Model] entries. - This would have a Direction keyword. - There could be Direction "I/O". - Arpad: Is the single DLL the only technical hurdle? - Michael M: That would resolve how many AMI files there are. - There might be other consistency issues. - Arpad: What if there is a single DLL? - For single DLLs a Direction parameter would have to be passed in. - Arpad: We should fix those issues if they are all that holds it back. - Ambrish: I don't think there are serious issues. - Fangyi: Why not make it simple? - Mike: The complication is only for EDA tools. - Fangyi: It is about making the specification simpler. - Arpad: The number of files on disk would be reduced. - Radek: It is a complication. - Michael M: It depends on where the adjudication is handled. - Michael M: I can't proceed until this is resolved. - Do we allow a single DLL for both TX and RX? - Arpad: We should all answer this. - Michael M: I may make a motion at some point. - Arpad: There should be an email. AR: Michael M send list email about directionality issues. ------------- Next meeting: 17 Mar 2015 12:00pm PT ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives